You will find below some hints for the EMBnet course on pairwise sequence comparison.

For most of the tools used during this practical, you can use the ID or Accession instead of copy & pasting the sequence.
Make sure you specify the right input sequence format in the pull-down list (i.e. SwissProt ID or AC).

• By default, you will get a local alignment. This means that a subsequence of the first sequence is aligned with a subsequence of the second sequence.
  
• A global alignment would align both sequences from beginning to end, whereas a local alignment can begin and end anywhere inside both sequences.
• It is often interesting to see not only the best local alignment, but also the ones with a lower score. Imagine for example 2 sequences:
  seq1 is composed of the domains A - B - B' - D
  seq2 is composed of the domains A - B - D
  
  A global alignment might produce the following result:
  

  seq1  A - B - B' - Dseq2  A - B   -    D

  
  A local alignment will also show you in its list, a local alignment of B' of seq1 and B of seq2, thusshowing the duplication and modification of the domain. It might show you something like:
  

  seq1  A - Bseq2  A - Bseq1  Dseq2  Dseq1  B'seq2  B

  
  Lalign allows you to specify the maximum number of local alignments you want to see.
• It is the percentage of aligned residues that are identical. It does not take into account similarity between amino-acids.
• The symbol ":" is used for identical aligned residues, whereas "." is used for similar aligned residues,i.e those that have a positive score in the scoring matrix used to compute the alignment.
  Be warned that those symbols may change in different programs.
• See the previous question. The blanks are used for aligned residues that have a score less or equal to 0 in the scoring matrix.Therefore the parameter influencing this is the scoring matrix.
• If the gap opening penalty is less negative than the gap extension penalty, the alignment will have many short gaps. On the contrary, ifthe gap opening penalty is more negative than the gap extension penalty there will be fewer but longer gaps.
• For the BLOSUM series of scoring matrices, the stringency increases with the number following the matrix name. I.e. the BLOSUM35 matrix willgive a good score even to distantly related amino-acids, whereas the BLOSUM80 matrix gives a positive score only to very closely related amino-acids.
  For the PAM series of scoring matrices it is the reverse: the higher the number the less stringent the matrix ;-)

• Look at the Dotlet documentation ;-)
• The gray level of a pixel represents the score of the window of aligned residues at this position in both sequences. The gray scaleon the right allows you to adjust how the colours map onto the scores.
• Most of the time an optimal position for the grayscale is, when it extends just across the smaller "bump" in the purple score distribution.Thus noise will be masked as black and the high scores are white.
• The diagonal lines represent alignments between the two sequences. By clicking inside the dotplot you can see the position of the two sequences againsteach other.
• It's not that difficult ;-)
• When the window size is reduced, the dotplot will have more background noise, but also more details. If the window size is increased, the details are smoothed out.
• This is usefull to identify repeats in a protein (two or more parallel diagonals) and low complexity regions (fuzzy spots).